Claude Code + Codex + Buzz is INSANE!

Today, we're going to be talking about agent operating systems, how to use them, how to get the most out of them, what you can build and automate with them, and some of my best workflows for automating, building anything with agent operating systems. And these are just powerful ways, basically link all your agents together with a memory system to get the most out of them so that you can have all your agents working together inside one space. Now we're going to be answering some questions from the AI Profit Boardroom, which is an amazing community where you can learn, grow, and scale with AI automation, lots of great people in there. And for example, we've got this question that we're about to answer.

So if you want to ask me questions like this and I'll make a video tutorial for you, feel free to post inside the AI Profit Boardroom and let's get straight into it. So Jessica is asking about Hermes agent, Hermes primary model and free fallback. So before there was actually Quen 3.6 available for free, which you could use as a free API with Hermes agent. But they've actually removed the beta trial period for that model, which means you can't use it for free.

However, don't worry, there's some other options. So what you could actually do is you could have as a free fallback, OmniRoot. This is a free way where you can code across like 90 different APIs and you can book that for free into your AI. So OmniRoot is an open source project.

It basically gives you like free models that you can use, but it automatically selects all of them and has like loads of different fallbacks automatically. So what that means, for example, is a lot of these free APIs get rate limited. So how do you fix that? Well, what you can do is you can use OmniRoot so that if one free API doesn't work or runs out or expires or gets rate limited, you can switch to a different one.

And if you want a full training on that, you can get that over here. So Quen 3.6 plus is not available anymore for free, but what you can use as well instead of this is there's actually two other new APIs and bear in mind like on NewsPortal with Hermes Agent, they're always releasing new APIs. So if you want a direct model instead, you can use Ling 3.0 Flash or free from Tencent Hanyan and to do that, all you do is you go into your AgentOS like so, then you're going to switch over to Hermes. From here, you're going to go to the manage section in your dashboard.

Then from the model section, you can actually change the model to OmniRoot or anything like that. Now, if you do get stuck, another option and you can see here that we can select NewsPortal, so if we scroll down, we can see all the free models available. There's actually four free models now, which is great or five actually. So Step 3.7 Flash, Poolside, Laguna XS 2.1, Ling 3.0, Laguna S 2.1 and then Tencent HY 3.3, so those are all different models that you can use as well.

So plenty of free options that you can use with Hermes Agent if you want to get the most out of it. Also, I noticed that you said you're going to use DeepSeek. I don't know what plan that comes with, but one thing that I found really good is also Kimi or Minimax. Kimi and Minimax are pretty good models for plugging in with your coding plan directly into Hermes Agent, which means that if you're already on the subscription for any of these models, you can just use the subscription instead of paying for the agent.

Now, we have another question as well about Obsidian and how to get the most out of that. So this is my memory system. And the great thing is, for example, when Quart Opus 5 just came out, we can just plug in context from day one with Obsidian into our AI agents. And if something new comes out, for example, like Buzz, then we could plug our memory system into Buzz and any other AI agent, even if it's not on the agent OS.

So how do you do this? Well, for example, let's say we were using something like Buzz. First thing we need to do is make sure that we have Obsidian set up, which we've got over here. And then you can actually ask your AI agents to export all of their memories and organize it into folders and files inside Obsidian.

And what the agent will do is start creating an Obsidian setup with all of your folders and files so that they're all neatly organized. So if you look at all of these, these are all different memories plugged in together, linked together so that our AI agents understand everything about us. It's what I call a memory galaxy. You see how we can visualize how our AI agents understand the context and the memory of everything that was done over the last few weeks.

So if we click on this link, for example, you can see it says LLM wiki. It explains one LLM wiki. Perfect. So what we can do from here is then we can go into any of our AI agents and say, connect to Obsidian locally and start using Obsidian for your main memory.

And that's basically it. We do have a more in-depth tutorial inside the AI Profitable Environment. It's called the Infinite Context Engine. And this basically uses a combination of Hermes agent with only an Obsidian, but you can replace Hermes with whichever agent you prefer.

You know, that could be Claude. It could be OpenCore. It could be Buzz. It could be Obsidian.

Whatever you, the main point is that you have a system that you can basically build and improve over time. And that becomes your competitive advantage because every day we use this, it improves, it gets better. We get more knowledge. The links become better.

All the information inside there becomes more useful for our AI agents as well. Now we had another question from Rajab who actually wants to learn keyword research, building blocks, how to do SEO and automate it all. Now, if you want to use our AgentOS system, we have that right here. And basically you can get this inside the AI Profitable Boarding and you can plug in a keyword, plug in a case study, and then deploy that to your website.

And to set it up, it's just a quick zip file that you give to your AI agents. And you ask them to set up with the SEO tool inside your AI agents. Now, if you have a look, for example, this is working really well for our websites here. So they're all on a really good trajectory, growing nicely over time with SEO.

And if you're wondering, okay, how do we train AI agents to do SEO for us? We have like a 12 page SEO skill that we just give to our SEO agents and they can start grading from there. So the simple version is you can use our skill for SEO and then give it to your AI agents and they can start creating content for you. The more complex option that's more powerful, but takes a bit longer to configure is you can use our AgentOS, plug in your websites with your Netlify access token, if you have your websites hosted on Netlify, or you can give them your WordPress login details with your API password, and they can log into your WordPress website and start creating content for you as well.

So two options right there. We had a question from Jessica, which was like, how do you wire your SEO into the AgentOS? So let me show you a quick example. So I actually set this up on a fresh setup this morning, just to show you.

So we attached the zip file into Claude and said, set this up for me and configure Claude so it works in the chat. Then it will ask you to log in and authorize your CLI with the AgentOS. Just tell it once you've done, and then it will have it set up and working for you like this. So pretty simple.

Step number one, attach the zip and say, set this up for me and configure Claude so it works. And you can do that for any CLI that you want to plug in there. And then step number two, tell it once you've logged in and then it's good to go. I'm actually going to build out a FAQ section on the AgentOS set up here.

So like, if you want to know how to configure CLIs, we've got examples. If you want to know how to set up Obsidian, we've got that. If you want to know how to set up a VBS, the AgentOS, you've got that right there. And it's just like all nearly organized for you to get.

And we've got a one hour video tutorial. If you want to go like really, really in depth on how to use AgentOS and get the most out of it. Some questions from, who says, Hey, awesome people. I'm looking to streamline how we build and share custom GPTs for each new client.

So these are some questions. So it's the first question for custom GPTs. And if you're not sure what a custom GPT is, here's an example. And specifically, Dr.

Fazalou is asking about setting up custom GPTs for clients, which is something that we do all the time. So you can see an example here. Can this workflow can come up with ideas for new videos, right? So how do we come up with those new ideas?

Well, basically we actually have a step-by-step process that we can give you for sharing custom GPTs. And the first question we got was, would you recommend hosting or self-hosting an open source NLM? So your 15 member team can access it. Personally, we don't do that.

We just have a shared company chat GPT account, and then my whole team can create custom GPTs there. Next question is, is it better to use chat GPTs, custom GPTs, despite the 20 file limit? I would say definitely yes, because number one, you probably won't need more than 20 files attached. We never do.

And then also you don't even need to share credentials. Like what you can actually do is you can just share the link to the custom GPT that you've created, and you just have one person on your team who creates the custom GPTs. Like for us, we just have two account managers who create custom GPTs for our clients. And that way we save time.

We keep it all accountable to one or two people and it keeps everything clean. And then the third question is, if file limits are a concern, what would you do? For me personally, we never find file limits are a concern. It's never really been an issue.

Great question. And that's basically it. We've answered all the questions inside the AI Profit Boardroom. If you want to ask me questions like this, feel free to post inside there.

And also there's so many great people inside there to connect with, but also to help you, right? There's loads of great wins. Lots of people sharing stuff. You know, four pages of testimonials and wins from people learning and growing and getting great results, which is awesome to see.

So it's just a really positive place where people can connect, learn. I think that learning AI as well is a long journey and we're all super early to it. So if you can start learning it now, number one, you're going to get a headstart. But number two, you can connect with some amazing people.

And these are all people trying to grow. These are all people trying to actively improve and learn and have fun along the way, which I think is fantastic. So if you want to connect with myself and everyone inside the group, feel free to get the AI Profit Boardroom. Link in the comments description, or just go to theaiprofitboardroom.com.

Inside the classroom, you also get access to all of our best trainings, our tutorials. We've got beginner to expert training over here. We have all of our step-by-step courses over here, new daily updates as well. And then you can also jump on four weekly coaching calls per week.

So you can get help four times a week, jump on a live call, share your screen, ask questions, et cetera. Inside the map, you can connect with people locally new who are building AI agents, just like you. And I'll see you inside there. Thanks for watching.